Solving Large-scale Eigenvalue Problems in SciDAC Applications

Slides:



Advertisements
Similar presentations
THE FINITE ELEMENT METHOD
Advertisements

Copyright 2011, Data Mining Research Laboratory Fast Sparse Matrix-Vector Multiplication on GPUs: Implications for Graph Mining Xintian Yang, Srinivasan.
A Large-Grained Parallel Algorithm for Nonlinear Eigenvalue Problems Using Complex Contour Integration Takeshi Amako, Yusaku Yamamoto and Shao-Liang Zhang.
05/11/2005 Carnegie Mellon School of Computer Science Aladdin Lamps 05 Combinatorial and algebraic tools for multigrid Yiannis Koutis Computer Science.
1 Modal methods for 3D heterogeneous neutronics core calculations using the mixed dual solver MINOS. Application to complex geometries and parallel processing.
A Discrete Adjoint-Based Approach for Optimization Problems on 3D Unstructured Meshes Dimitri J. Mavriplis Department of Mechanical Engineering University.
Applied Linear Algebra - in honor of Hans SchneiderMay 25, 2010 A Look-Back Technique of Restart for the GMRES(m) Method Akira IMAKURA † Tomohiro SOGABE.
Computational Science R&D for Electromagnetic Modeling: Recent Advances and Perspective to Extreme-Scale Lie-Quan Lee For SLAC Computational Team ComPASS.
Solving Linear Systems (Numerical Recipes, Chap 2)
COMPASS All-hands Meeting, Fermilab, Sept Scalable Solvers in Petascale Electromagnetic Simulation Lie-Quan (Rich) Lee, Volkan Akcelik, Ernesto.
Numerical Parallel Algorithms for Large-Scale Nanoelectronics Simulations using NESSIE Eric Polizzi, Ahmed Sameh Department of Computer Sciences, Purdue.
Modern iterative methods For basic iterative methods, converge linearly Modern iterative methods, converge faster –Krylov subspace method Steepest descent.
1cs542g-term Notes  In assignment 1, problem 2: smoothness = number of times differentiable.
CS 584. Review n Systems of equations and finite element methods are related.
Mar Numerical approach for large-scale Eigenvalue problems 1 Definition Why do we study it ? Is the Behavior system based or nodal based? What are.
Scientific Computing Seminar AMLS, Spectral Schur Complements and Iterative Computation of Eigenvalues * C. BekasY. Saad Comp. Science & Engineering Dept.
Advancing Computational Science Research for Accelerator Design and Optimization Accelerator Science and Technology - SLAC, LBNL, LLNL, SNL, UT Austin,
An introduction to iterative projection methods Eigenvalue problems Luiza Bondar the 23 rd of November th Seminar.
Large-Scale Density Functional Calculations James E. Raynolds, College of Nanoscale Science and Engineering Lenore R. Mullin, College of Computing and.
MATH 685/ CSI 700/ OR 682 Lecture Notes Lecture 6. Eigenvalue problems.
ECE 530 – Analysis Techniques for Large-Scale Electrical Systems Prof. Hao Zhu Dept. of Electrical and Computer Engineering University of Illinois at Urbana-Champaign.
Introduction The central problems of Linear Algebra are to study the properties of matrices and to investigate the solutions of systems of linear equations.
Linear Algebra and Image Processing
Solution of Eigenproblem of Non-Proportional Damping Systems by Lanczos Method In-Won Lee, Professor, PE In-Won Lee, Professor, PE Structural Dynamics.
Physics “Advanced Electronic Structure” Pseudopotentials Contents: 1. Plane Wave Representation 2. Solution for Weak Periodic Potential 3. Solution.
1 A Domain Decomposition Analysis of a Nonlinear Magnetostatic Problem with 100 Million Degrees of Freedom H.KANAYAMA *, M.Ogino *, S.Sugimoto ** and J.Zhao.
Algorithms for a large sparse nonlinear eigenvalue problem Yusaku Yamamoto Dept. of Computational Science & Engineering Nagoya University.
Qualifier Exam in HPC February 10 th, Quasi-Newton methods Alexandru Cioaca.
ParCFD Parallel computation of pollutant dispersion in industrial sites Julien Montagnier Marc Buffat David Guibert.
CFD Lab - Department of Engineering - University of Liverpool Ken Badcock & Mark Woodgate Department of Engineering University of Liverpool Liverpool L69.
Efficient Integration of Large Stiff Systems of ODEs Using Exponential Integrators M. Tokman, M. Tokman, University of California, Merced 2 hrs 1.5 hrs.
The Finite Element Method A Practical Course
In-Won Lee, Professor, PE In-Won Lee, Professor, PE Structural Dynamics & Vibration Control Lab. Structural Dynamics & Vibration Control Lab. Korea Advanced.
1 Incorporating Iterative Refinement with Sparse Cholesky April 2007 Doron Pearl.
Ab-initio Calculations of Microscopic Structure of Nuclei James P. Vary, Iowa State University Esmond G. Ng, Berkeley Lab Masha Sosonkina, Ames Lab April.
Computational Aspects of Multi-scale Modeling Ahmed Sameh, Ananth Grama Computing Research Institute Purdue University.
* 김 만철, 정 형조, 박 선규, 이 인원 * 김 만철, 정 형조, 박 선규, 이 인원 구조동역학 및 진동제어 연구실 구조동역학 및 진동제어 연구실 한국과학기술원 토목공학과 중복 또는 근접 고유치를 갖는 비비례 감쇠 구조물의 자유진동 해석 1998 한국전산구조공학회 가을.
On implicit-factorization block preconditioners Sue Dollar 1,2, Nick Gould 3, Wil Schilders 2,4 and Andy Wathen 1 1 Oxford University Computing Laboratory,
Quality of model and Error Analysis in Variational Data Assimilation François-Xavier LE DIMET Victor SHUTYAEV Université Joseph Fourier+INRIA Projet IDOPT,
Case Study in Computational Science & Engineering - Lecture 5 1 Iterative Solution of Linear Systems Jacobi Method while not converged do { }
*Man-Cheol Kim, Hyung-Jo Jung and In-Won Lee *Man-Cheol Kim, Hyung-Jo Jung and In-Won Lee Structural Dynamics & Vibration Control Lab. Structural Dynamics.
Al Parker July 19, 2011 Polynomial Accelerated Iterative Sampling of Normal Distributions.
23/5/20051 ICCS congres, Atlanta, USA May 23, 2005 The Deflation Accelerated Schwarz Method for CFD C. Vuik Delft University of Technology
James Brown, Tucker Carrington Jr. Computing vibrational energies with phase-space localized functions and an iterative eigensolver.
Report from LBNL TOPS Meeting TOPS/ – 2Investigators  Staff Members:  Parry Husbands  Sherry Li  Osni Marques  Esmond G. Ng 
Consider Preconditioning – Basic Principles Basic Idea: is to use Krylov subspace method (CG, GMRES, MINRES …) on a modified system such as The matrix.
F. Fairag, H Tawfiq and M. Al-Shahrani Department of Math & Stat Department of Mathematics and Statistics, KFUPM. Nov 6, 2013 Preconditioning Technique.
1 Instituto Tecnológico de Aeronáutica Prof. Maurício Vicente Donadon AE-256 NUMERICAL METHODS IN APPLIED STRUCTURAL MECHANICS Lecture notes: Prof. Maurício.
A Parallel Hierarchical Solver for the Poisson Equation Seung Lee Deparment of Mechanical Engineering
Multipole-Based Preconditioners for Sparse Linear Systems. Ananth Grama Purdue University. Supported by the National Science Foundation.
Finding Rightmost Eigenvalues of Large, Sparse, Nonsymmetric Parameterized Eigenvalue Problems Minghao Wu AMSC Program Advisor: Dr. Howard.
MA237: Linear Algebra I Chapters 1 and 2: What have we learned?
Conjugate gradient iteration One matrix-vector multiplication per iteration Two vector dot products per iteration Four n-vectors of working storage x 0.
UT-BATTELLE New method for modeling acoustic waves in plates A powerful boundary element method is developed for plate geometry The new method achieves.
Computational Fluid Dynamics Lecture II Numerical Methods and Criteria for CFD Dr. Ugur GUVEN Professor of Aerospace Engineering.
Parallel Direct Methods for Sparse Linear Systems
Hui Liu University of Calgary
G. Cheng, R. Rimmer, H. Wang (Jefferson Lab, Newport News, VA, USA)
Introduction The central problems of Linear Algebra are to study the properties of matrices and to investigate the solutions of systems of linear equations.
Challenges in Electromagnetic Modeling Scalable Solvers
Numerical Modeling for Semiconductor Quantum Dot Molecule Based on the Current Spin Density Functional Theory Jinn-Liang Liu Department of Applied.
Introduction to Multigrid Method
1C9 Design for seismic and climate changes
Deflated Conjugate Gradient Method
L Ge, L Lee, A. Candel, C Ng, K Ko, SLAC
Supported by the National Science Foundation.
Conjugate Gradient Method
LARGE SCALE SHAPE OPTIMIZATION FOR ACCELERATOR CAVITIES*
Numerical Linear Algebra
RKPACK A numerical package for solving large eigenproblems
Presentation transcript:

Solving Large-scale Eigenvalue Problems in SciDAC Applications Chao Yang Lawrence Berkeley National Laboratory June 27, 2005

People Involved LBNL: W. Gao, P. Husbands, X. S. Li, E. Ng, C. Yang (TOPS) J. Meza, L. W. Wang, C. Yang (Nano-science) SLAC: L. Lee, K. Ko Stanford: G. Golub UC-Davis Z. Bai

SciDAC Applications Accelerator Modeling Nano-science

Algorithms Krylov Subspace Method Alternatives Optimization based approach non-linear solver based approach Multi-level Sub-structuring Non-linear Eigenvalue Problems Structure preserving methods Optimization based method

Krylov Subspace Method Widely used, relatively well understood (Polynomial approximation theory): Convergence of KSM: Well separated, large eigenvalues converge rapidly the starting vector

Acceleration Techniques Implicit Restart Spectral transformation ARPACK filter out unwanted spectral components from v0

Using KSM in accelerator modeling the spectrum of the problem Example: H60VG3 structure, linear element, N=30M, nnz=484M 1024 CPUs, 738GB Ordering time: 4143s Numerical Factorization: 133s Total: 5068s for 12 eigenvalues Software: PARPACK (implicit restart) + SuperLU, WSMP (spectral transformation)

Limitations of the KSM High degree polynomial needed for computing small clustered eigenvalues many matrix vector multiplications Spectral transformation can be expensive memory limitation scalability Not easy to introduce a preconditioner eigenvectors of P-1A are different from eigenvectors of A

Alternative algorithms Optimization based approach Minimizing Rayleigh Quotient Minimizing Residual (Wood & Zunger 85, Jia 97) Nonlinear equation solver based approach (Jacobi-Davidson) Newton correction Preconditioner stopping criteria for the inner iteration (Notay 2002, Stathopoulos 2005) Allows us to solve problems with more than 90M DOF

Multi-level Sub-structuring (for computing many eigenpairs) Domain Decomposition concept Multi-level extension of the Component Mode Synthesis (CMS) method (Bennighof 92) Decomposition can be done algebraically (Lehoucq & Bennighof 2002) Success story in structure engineering.... Error analysis Extend to accelerator modeling

Single-level Sub-structuring Matrix Partition Block elimination Sub-structure calculation (mode selection) Subspace assembling

Mode Selection

Implementation & cost attractive when: 1) the problem is large enough Flops: more than a single sparse Cholesky factorization Storage: Block Cholesky factor + Projected matrix + some other stuff NO triangular solves (involving the original K and M), NO orthogonalization attractive when: 1) the problem is large enough 2) a large number of eigenvalues are needed

AMLS vs. Shift-invert Lanczos (SIL) DOF=65K, 3 levels of partition

Cavity with External Coupling Open Cavity Waveguide BC Vector wave equation with waveguide boundary conditions can be modeled by a non-linear eigenvalue problem Closed cavity is an approximation for the real model. Usually the cavity has some sort of openings. For example, in accelearator cavity, we need put power into cavity and damp high-order-modes. These external coupling requires waveguide connected to cavity. Waveguide shape can be circular, rectangular or coax. To model the open cavtity with waveguides, we can introduce a virtual BC on waveguide as shown. The problem is emerged from a cavity with external coupling through waveguides. We can put waveguide boundary conditions on the virtual boundaries. Curl-curl equation along with waveguide boundary conditions will be discretized using finite-element methods (Electric field E is expanded by a set of vector basis functions.) and becomes a non-linear eigenvalue problem, where matrix K corresponds to curl-curl operator. M is mass matrix. Matrix W is waveguide matrix and has nonzero entries only on waveguide boundary. Note that k_cj is a physical known quantity called cutoff. With

Quadratic Eigenvalue Problem Consider only one mode propagating in the waveguides Algorithms Linearize then solve by KSM (does not preserve the structure of the problem) Second Order Arnoldi Iteration (Bai & Su 2005) project the QEP into 2nd order Krylov Subspace

Second-Order Krylov Space (Bai)

SOAR is faster and more accurate (than linearization) Accelerating cavity model for international linear collider (ILC) 9-cell superconducting cavity coupled to one input coupler and two Higher-Order-Mode couplers. NDOFs=3.2million, NCPUs=768, Memory=300GB 18 eigenpairs in 2634 seconds (linearization took more than 1 hour)

Electronic Structure Calculation Etotal(X) = Ekinetic + Eionic + EHartree + Exc wave function n – real space grid size, e.g. 323~32000 k – number of occupied states, 1~10% of n Charge density Ekinetic = Eionic = EHartree= Exc =

Non-linear Eigenvalue Problem Total energy minimization KKT condition

The Self Consistent Field Iteration Input: initial guess and Output: Major steps For i=1,2,…,until converged Form Compute k smallest eigpairs of Consistent (try tex point)

Direct Constrained Minimization (DCM) For i=1,2,… until convergence Form Compute If (i>1) then set else Solve Be consistent with subscripts etc K is the preconditioner

DCM vs. SCF Atomic system: SiH4 Discretization: spectral method with plane wave basis: n=323 in real space, N=2103 (# of basis functions) in frequency space Number of occupied states: k = 4 PETOT version of SCF uses 10 PCG steps (inner iterations) per outer iteration DCM: 3 inner iterations

Concluding Remarks Krylov Subspace Method (with appropriate acceleration strategies) continues to play an important role in solving SciDAC eigenvalue problems Steady progress has been made in alternative approaches that can make better use of preconditioners Multi-level sub-structuring is promising for computing many eigenpairs Significant progress made in solving QEP Non-linear eigenvalue problems remain challenging